AI Tooling Engineer

• Build the platform with the Principal Engineer in Tooling: new capabilities every week, from design to production.
• Connect new internal systems and SaaS tools to the platform, including APIs, data, shared models, and access permissions.
• Build agents and their tools, including MCP servers and the context needed for high-quality answers.
• Own evaluation sets, benchmarks, tracing, cost per answer, and issue resolution.
• Turn proofs of concept into secure, production-ready internal cloud tools.
• Use AI coding agents with specifications, tests, and reviews to maintain code quality.
• Contribute to company-wide AI tooling initiatives, including tool evaluation and standardization, cost and usage tracking, and standards for AI-built internal applications.
• Help other teams build on the platform.
• Engage platform users across the company to identify priorities and explain delivered work to technical and non-technical audiences.

• 3+ years building and shipping production software, mostly in Python, with relational databases.
• Has built and shipped with LangGraph at least one agent that real people use: tool calling, retrieval and embeddings, prompt and context design.
• Hands-on experience with LLM APIs from the main providers, cost and token optimisation.
• Hands-on with MCP or similar tool protocols, as a server and as a client.
• Evaluates AI systems properly: test sets, LLM-as-judge, regressions when a prompt or a model changes; tracing and observability with Langfuse or similar.
• Integrates APIs and data: SaaS APIs, SQL, data warehouses.
• Experience with cloud infrastructure (Google Cloud and/or AWS): containers, managed databases, identity and access, CI/CD.
• Uses AI coding agents every day and knows how to keep their output in check.
• A sense for security and permissions: identity, least privilege, people see only what they are allowed to see.
• Comfortable with version control (GitHub/Bitbucket).
• Excellent written and spoken English, technical and non-technical.
• Familiarity with web data or web scraping.
• Django and Postgres, including vector search with pgvector.
• Enough frontend to build and change a chat interface (React, Next.js).
• Has built internal tools that colleagues adopted.
• Semantic layers or knowledge graphs.
• Memory for agents, multi-agent setups.
• Previous remote/distributed team experience.

• Become part of a self-motivated, progressive, multi-cultural team.
• Freedom and flexibility to work remotely.
• Work with cutting-edge, open-source technologies and tools — and shape the AI tooling the rest of the company relies on.